(copy and paste error.)

]]>I don’t understand why “\epsilon_{s}” () is in the following

\displaystyle T_{tt}^{4} \approx T_{t}^{4} + \frac{\epsilon_{s} R \sigma T_{t}^{4} }{\epsilon OLR}= T_{t}^{4}[1+\frac{\epsilon_{s}R}{\epsilon OLR}]

In the partial derivative

\displaystyle \frac{\partial [T_{tt}/T_{t}]}{\partial \epsilon} = – \frac{\epsilon_{s} R T_{t}}{4 \epsilon^{2} OLR} [1+\frac{\epsilon_{s}R}{\epsilon OLR}]^{-0.75}

Why is T_{t} () included in

\epsilon_{s} R T_{t}

Perhaps you intended to take the partial of just the new temperature instead of the ratio.

Assuming that, it appears that as epsilon gets smaller the partial derivative gets more negative, which implies that increasing CO2 should make the stratosphere warmer. Needless to say, I am very confused.

**Response: Adding CO2 increases absorptivity (and hence emissivity…if I understand your question right). So it looks like the sign is right…but yes, I fixed a couple mistakes in those equations. The dimensions didn’t even make sense before. Thanks. This didn’t affect any of the graphics though – chris**

25 years ago, under much lower CO2 conditions, there was a very bad cold snap in Louisiana, but even that wasn’t cold enough to freeze over the local creek…which a 10 years old could jump across.

How could a net difference of just 0.26W freeze over the Mississippi and the Thames?

**Response: I agree the solar forcing is pretty small. That leaves “indirect” amplification mechanisms, other forcings (volcanic), and/or internal variability, particularly at the very regional level. I’m not convinced that climate responses to changes in the sun in the last millennium are outside the range of noise you’d get from just running a large ensemble of simulations at this regional scale, maybe even globally **

Instead of 1.5W they calculated a difference of 3.3W in either 1995 or 1997, and then in 2000 recalculated it as 2.8W.

Also, given the information in the paper, it seems they are not using the same dataset as you, for example all of their solar irradiance data are several watts higher anyway.

**Response: Some of the older papers (e.g Lean, Hoyt and Schatten) aren’t considered valid anymore by the community that does these solar reconstructions. There is a bit of discussion here and the Schmidt et al. paper I referenced discusses the new reconstructions which are what is being used now in the current installment of last millennium climate simulations. The older Lean 2000 result was based on another paper inferring Maunder Minimum values with comparison of solar activity with “sun-like” stars, but there’s other lines of evidence and issues the statistical robustness of their sample. But you’ll have to follow the literature for details… **

You write “The typical solar cycle amplitude is on the order of ~0.17 W/m2 and the difference from Maunder Minimum to present is around 1.5 W/m2 or so” – and then explain conversion of TSI to forcing, but the first of those figures is already converted, right?

**Response: You’re correct, I changed the value to make it consistent (TSI amplitude). -chris**